MISP Guard, a proxy enforcing security policies on MISP feed data, has a vulnerability where it incompletely validates required taxonomy tags. The guard verifies only that at least one tag from a required taxonomy is allowed but fails to ensure all tags from that taxonomy are allowed. This flaw allows events or attributes to carry both allowed and non-allowed tags simultaneously, bypassing data classification restrictions and potentially exposing sensitive data to unauthorized feed consumers. The issue affects event-level and attribute-level tag validation, including nested and shadow attributes. An attacker able to modify MISP events in a feed source can exploit this to leak higher-classification data through feeds intended for lower-classification data. The vulnerability has a CVSS 4.0 base score of 6.1 (medium severity). No patch or fix information is provided.
The strapi/strapi Core Admin / Webhook Runner component is vulnerable to a semi-blind Server-Side Request Forgery (SSRF) via webhook follow-redirect behavior. The webhook dispatcher follows HTTP redirects automatically without verifying the IP address after redirection or checking if the target is within private network ranges. Additionally, non-2xx HTTP responses from target endpoints are reflected back to the caller.
The strapi/content-manager package has a vulnerability in its content manager preview URL endpoint where content-scope authorization is missing. This flaw allows low-privileged administrators to generate preview URLs without needing to know document identifiers or having read permissions.
A vulnerability in Thunderbird allowed clicking on news: or snews: links referencing a message-ID to cause the client to connect automatically to an attacker-specified host and port without user consent, even if the server was not configured. Line breaks in the message-ID were sent unmodified, enabling crafted links to send arbitrary commands to reachable services. Additionally, snews: links connected without TLS, exposing data to interception. This issue was fixed in Thunderbird versions 159 and 153.6.
A vulnerability in Thunderbird on Windows allows bypassing a security check that prevents mailbox: URLs from accessing files on remote Windows network shares by using extended-length UNC paths. This can cause Thunderbird to connect to an attacker-controlled SMB server, leaking the user's Windows account name and NTLM authentication data. The leaked NTLM data could be used for offline password cracking or relay attacks. This issue affects only Thunderbird on Windows and was fixed in Thunderbird versions 159 and 153.6.
Thunderbird versions prior to 159 and 153.6 accepted calendar locations from directory lookups that could point to local files or remote network shares. This flaw allowed a malicious server or network attacker to cause Thunderbird to subscribe to such locations, potentially leaking Windows NTLM credentials on every calendar refresh and enabling unauthorized writes to local or remote files.
A vulnerability in Thunderbird allowed automatic loading of attacker-controlled URLs via OpenPGP key attachments without user interaction, bypassing remote content blocking. On Windows, this could leak NTLM credentials to an attacker-controlled network share. The issue was fixed in Thunderbird 159 and 153.6.
The strapi/review-workflows package contains a broken access control vulnerability that allows updates to be made without verifying if the caller has the appropriate update permissions for the underlying document.
CVE-2026-117988 is a sandbox escape vulnerability caused by incorrect boundary conditions in the Graphics: CanvasWebGL component. This flaw allows an attacker to break out of the sandbox environment. The issue was fixed in Firefox 159, Firefox ESR 153.6, Thunderbird 159, and Thunderbird 153.6.
CVE-2026-117989 is a use-after-free vulnerability in the Graphics component of Mozilla products. This security flaw was addressed and fixed in Firefox 159, Firefox ESR 153.6, Thunderbird 159, and Thunderbird 153.6. The vulnerability is rated with high impact by the vendor. No known exploits in the wild have been reported.
CVE-2026-117990 is a high-impact vulnerability in the Graphics: WebRender component of Mozilla Firefox and Thunderbird. It allows sandbox escape due to incorrect boundary conditions. This flaw was addressed and fixed in Firefox 159, Firefox ESR 153.6, Thunderbird 159, and Thunderbird 153.6.
CVE-2026-117945 is a vulnerability involving incorrect boundary conditions in the XML component of Mozilla products. This flaw was addressed and fixed in Firefox 159, Firefox ESR 153.6, Thunderbird 159, and Thunderbird 153.6. The vulnerability is considered high impact by Mozilla. No known exploits in the wild have been reported.
CVE-2026-57191 is a high-severity classic stack buffer overflow vulnerability in Asterisk's res_pjsip_pubsub module. It occurs due to an unbounded sscanf call when parsing the MWI NOTIFY Message-Account field. An unauthenticated attacker can send a crafted packet to trigger this overflow, corrupting the transport and disabling all SIP functionality until the service is restarted. Fixed versions are available.
MLflow tracking servers running without authentication or with basic authentication are vulnerable to an arbitrary file read vulnerability via crafted model versions tagged as prompts with schemeless local paths. This allows remote attackers to read arbitrary files on the server filesystem with the privileges of the MLflow server process. The vulnerability affects MLflow versions 3.10.0 through 3.16.1 and is fixed in version 3.17.0.
MLflow's CrewAI autologging feature in versions >=2.19.0 and <3.17.0 leaks CrewAI LLM provider API keys in trace telemetry. The vulnerability arises because MLflow serializes the CrewAI Agent object's __dict__ including the LLM object, whose string representation contains the API key in plaintext. This sensitive information is unintentionally sent to MLflow's trace backend or any configured OTLP destination without explicit logging by the application. The issue affects MLflow 3.16.0 with CrewAI 1.15.18 and is fixed in MLflow 3.17.0.
MLflow versions from 3.12.0rc0 up to but not including 3.17.0 are vulnerable to unauthenticated remote code execution via a flaw in third-party scorer deserialization. The vulnerability arises because the server imports and instantiates Python classes from request-controlled module paths under allowed package prefixes, which can be exploited by uploading malicious Python modules into writable artifact directories within these allowed packages. This leads to arbitrary code execution with the privileges of the MLflow server process, which by default runs as root in the official image. The issue is critical and has a CVSS score of 9.8. A fix is available in MLflow 3.17.0.
CVE-2026-57192 is a medium severity vulnerability in Asterisk's Codec2 decoder that allows an authenticated attacker to cause a heap buffer overflow via an out-of-bounds write. The flaw arises from a floor/ceil sample-count mismatch when the codec_codec2 module is loaded and Codec2 is configured on the endpoint. Exploitation requires SIP authentication and establishing a call using Codec2 with crafted audio packets. Fixed versions include Asterisk 20.20.1, 21.12.3, 22.10.1, 23.4.1, and Certified Asterisk 20.7-cert11, 22.8-cert3.
A vulnerability in the Asterisk telephony toolkit's chan_ooh323 channel driver allows malformed Q.931 elements in an OOH323 request to crash the system. This issue exists because a prior fix (CVE-2022-37325) was not forward-ported to versions after Asterisk 20. The vulnerability affects versions where chan_ooh323 is explicitly compiled and publicly reachable. Fixed versions include 21.12.3, 22.10.1, 23.4.1, and Certified Asterisk 22.8-cert3.
CVE-2026-57200 is a path traversal vulnerability in Asterisk affecting versions prior to 20.20.1, 21.12.3, 22.10.1, 23.4.1, and Certified Asterisk 22.8-cert3. An authenticated attacker with read-only Asterisk REST Interface (ARI) credentials may exploit this to perform write operations via REST-over-WebSocket, leading to arbitrary module path loading and potential remote code execution. The vulnerability requires the Asterisk HTTP server to be enabled and reachable, which is disabled and bound to localhost by default. This issue has been fixed in the specified versions.
A ReDoS vulnerability exists in dd-trace's URL query string obfuscation regex, causing excessive CPU usage when processing crafted inputs. This can stall request processing and lead to denial of service in applications using affected HTTP tracing integrations. The issue affects dd-trace versions >=2.0.0 <5.131.1 and >=6.0.0 <6.20.1. A fixed version is available in 5.131.1 and 6.20.1. A workaround involves setting an updated regex pattern via DD_TRACE_OBFUSCATION_QUERY_STRING_REGEXP and restarting the application.